RFR: 8341977: Replace predicate walking and cloning code for Loop Peeling with a predicate visitor [v2]

[Christian Hagedorn]

On Mon, 28 Oct 2024 10:11:21 GMT, Emanuel Peter <epeter at openjdk.org> wrote:

>> src/hotspot/share/opto/predicates.hpp line 952:
>> 
>>> 950:   Node* _new_control;
>>> 951:   PhaseIdealLoop* const _phase;
>>> 952:   ParsePredicateSuccessProj* _current_parse_predicate;
>> 
>> It looks to me like this could be a boolean, correct?
>
> Then the name could also be more descriptive

I changed it into a `bool`. The idea is to simulate the old behavior that we only create the Assertion Predicates if there are Parse Predicates available. I don't think this is generally correct - we could still split loops later when Parse Predicates have already been removed. But I want to fix this at other places as well and thus defer this to a separate change later.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/21679#discussion_r1819024953